Multiband galaxy morphologies for CLASH: a convolutional neural network transferred from CANDELS

M. Pérez-Carrasco, G. Cabrera-Vives, M. Martinez-Marin, P. Cerulo, R. Demarco, P. Protopapas, J. Godoy, and M. Huertas-Company.

During my undergraduate studies, I gained my first experience with machine learning while working on this project. Specifically, I joined the astroinformatics group at the University of Concepción to develop a model for classifying galaxy morphology. Our project's goal was to classify galaxies in different photometric bands from the CLASH survey. We began by training the model with galaxies from CANDELS and then transferred this knowledge to the CLASH galaxies. This experience has not only motivated but also compelled me to pursue additional data science courses so I can continue to apply machine learning techniques in astronomy.

We presented this work at the "Big Data Astronomy Workshop"

March 1-2, 2018

Faculty of Engineering

University of Concepción